Eye-Tracking Dataset: Emoji Keyboard Gaze Search (Raw and Processed)

# Raw and Processed Eye-Tracking Data: Emoji Keyboard Gaze Search Study ## Folder Structure raw_data/ [Task Type]/ # "Scenario-based Selection" or "Target Search" [layout]/ # "android" (vertical) or "ios" (horizontal) [participant_id]/ extracted_data.xlsx # 6 sheets, one per task (Task1-6) [layout]-scroll/ [participant_id]/ [task_number].txt # scroll position log per task processed_data/ [Task Type]_[layout]_p[N]_task[N].csv # 684 files, episode-level merged data ## Raw Eyetracker Data (extracted_data.xlsx) Column descriptions:- Id: Sample index- Time(s): Timestamp in seconds from task onset- X(%), Y(%): Gaze position as percentage of screen- X(px), Y(px): Gaze position in pixels- Eye State Type: 注视 (fixation) or 眼跳 (saccade), classified via I-VT algorithm (velocity threshold 30°/s)- Velocity(degree/s): Angular velocity of eye movement ## Raw Scroll Data ([layout]-scroll/[participant]/[task].txt) First line: total task duration (总时间: XX.XXXXs)Subsequent lines: scroll position values, recorded from task onset to task end.- For vertical (android) layout: values correspond to Y-axis scroll offset- For horizontal (ios) layout: values correspond to X-axis scroll offset- Sampling rate: [TBD - 확인 필요] ## Processed Data (episode-level) Generated by merging raw eyetracker and scroll data (synchronized by time index/timestamp), then applying episode segmentation (new episode defined when cumulative displacement between consecutive fixations exceeds 30px). Columns: episode_id, eye_state_type, x(px), y(px), time_start, time_end, duration, velocity_min, velocity_max, velocity_mean, velocity_sd